Automatic Pouring Robot. Akilah Harris-Williams Adam Olmstead Philip Pratt-Szeliga Will Roantree

Transcription

1 Automatic Pouring Robot Akilah Harris-Williams Adam Olmstead Philip Pratt-Szeliga Will Roantree

2 Overview Objective and Motivation Mechanical System Modeling, Simulation and Verification Tilt Pan Pouring Trajectory Generation Control Design Tilt Pan Webcam Image Processing Uncertainty Analysis

3 Objective and Motivation To develop an Automatic Pouring Robot Industrial Applications Casting Robot Hazardous Materials Robot

4 Purpose Smoothly pour water from a bottle into a cup without spilling Consists of a bottle, a pan/tilt mechanism, a computer controller, a cup and a webcam The bottle will be mounted on the mechanism Computer controller will rotate the bottle towards the cup Webcam will view the height of water in the cup.

5 Mechanical System

6 Tilt Modeling Mass Properties change with tilt angle and water volume. System dynamics are variable. Visual Basic program extracts Center of Mass, Inertial Tensor and Mass from Solidworks API at varying angles and volumes.

7 Tilt Friction/Gravity Model Upper Level System Tilt Dynamics Equation of Motion

8 Tilt Friction/Gravity Results Experimental system response versus Simulated system response to force of gravity.

9 Tilt Friction/Gravity Cancellation

10 Pan Modeling θ + a1 θ + a 2 sgn( θ ) = a 3V Used full parameter identification method Chirp input a1= a2= a3= rad Steady-State Error

11 Basic Trajectory

12 Basic Trajectories

13 Polynomial Trajectories

14 Basic Return Trajectories

15 Return Trajectories

16

17

18

19 Tilt Control Needs to follow a smooth trajectory Specifications: M o 0%, t r 0.75 s, e ss 1% Used PID controller System behaved similarly to model Linear model assumes constant volume Due to modeling error, needed an easily tweaked controller

20 Tilt Controller Simulated Simulated M o = 2.87% t r =0.327 s e ss =0.81% Actual M o = -0.13% t r =0.104 s e ss =0.138%

21 Tilt Controller - Actual Simulation was run assuming a constant volume Values will change due to: Constantly changing amount of liquid Any modeling error Values tweaked from simulation to dampen overshoot

22 Tilt Controller on Actual Trajectory

23 Pan Control Used for Disturbance Rejection Needs to have a quick rise time and low steady-state error Specifications M o 5%, t r 0.4 s, e ss 2% Lead-Lag Controller D ( s) = 30 ( s )( s ) ( s )( s + 0.9)

24 Pan Controller - Design

25 Pan Controller - Simulation Simulated: M o = % t r =0.336 s e ss =0.037 % Actual: M o = 5.23 % t r = s e ss = %

26 Tweaked Pan Controller Actual: M o =1.55 % t r =0.134s e ss =0.016 % t s =0.356s (2%) t s =1.548s (1%)

27

28 Sensor Subsystem - Webcam Objective: Pour a specified amount of liquid into a cup Need to know when to stop pouring to get the desired amount Our solution: Webcam

29 Image Processing Algorithm Want to find the volume of liquid in the cup Find highest pixel row occupied by the colored liquid (determined by green and blue thresholds) Correlate these pixels to actual volumes via experimentation

30 Webcam Timing Average frame rate with processing is 13 fps. When the volume in the cup approaches a certain percentage of the desired volume, MATLAB triggers xpc target to stop pouring. Percentage at which the trigger is sent is a function of: Return trajectory Communication delay between MATLAB and xpc Target Flow rate before stopping

31 Webcam Data of Sample Pour Stop Triggered

32 Uncertainty Analysis Modeling error Tilt: Friction Identification Pan: Steady-State Error Webcam error Noisy data Slow feedback Other sources of error Occasional encoder spikes

33 Modeling Error Tilt Friction ID Velocities never converge to a steady-state value, due to water sloshing in the bottle Different values in each direction Average was taken

34 Modeling Error - Pan Identified parameters work fairly well with data used to generate them. e ss of only 0.08 rad = 4.6 Does not work as well for all inputs

35 Assessment Mechanical System Sturdy apparatus key to entire project Model of Dynamic Tilt Parameters Works well after initial calibration of mass Webcam Image Processing Good response but prone to inaccuracies Disturbance Rejection Functions well, but with large settling time for large disturbances Pouring Control Occasionally spills due to stationary cup Overall Success!

36 Conclusion Areas of Improvement Third degree of motion Faster and more precise cup volume feedback Implement fuzzy logic controller based on constant webcam data Improve models and parameter identification

37